unexpected prematurely closed connection

[zffocussss]

Hi team:
we are using granian to run an asgi service which have two instances in AKS runtime,and there is two kong instances running before the granian http server.From time to time,there were some similar issues in kong side.

upstream prematurely closed connection while reading response header from upstream

it is proven that the granian worker closed the connection unexpectedly.but I could not find any footprint in the logs in granian side.How can it be fixed?I am not sure if it is a bug or intended.

granian --interface asgi main:app --host 0.0.0.0 --port 8002 --workers ${CPU_NUMBER} --respawn-interval 30 --workers-max-rss 1000 --workers-kill-timeout 120 --log-level info --metrics --metrics-address 0.0.0.0 --metrics-port 9090  --metrics-scrape-interval 10

my environment:
AKS 1.35
granian 2.7.9
kong 3.9.3